Daily Summary of Cybersecurity · Live edition, updated continuouslySunday 22 November 2026 · UTC
Latest Cybersecurity News Today: 22 November 2026
Bell and Cleverbridge breaches hit business data and cards as Signal blocks AI assistants
Bell Canada confirmed a cyberattack that stole enterprise customer data, while Cleverbridge said hackers took payment card numbers and CVV codes from its platform. Signal added protections to block Siri and Gemini from capturing chats on iOS and Android. Optimizely malware hit 178 organizations, and a Ghostscript patch fixed a sandbox escape.
Bell Canada says cyberattack stole enterprise customer data
Bell Canada said a cyberattack stole enterprise customer data after unauthorized access to multiple systems, including a pre-production system for network connections. Exposed data may include business contact details, ticket titles and specific IP addresses, but not login credentials or highly sensitive personal information. Bell called the actor agentic, suggesting AI tools may have been used.
Cleverbridge says hackers stole card numbers and CVV codes
Cleverbridge, an e-commerce platform for software and SaaS companies, said hackers copied payment card data from its platform after detecting unauthorized activity on November 4, 2026. A breach notification says the stolen data included card numbers, expiration dates and CVV/CVC codes, while no other personal data was copied. The company said the incident was not ransomware.
Signal blocks Siri and Gemini from capturing chats on iOS and Android
Signal added protections that block screenshots on iOS 27 and prevent AI assistants from capturing conversations on supported Android devices, targeting Siri and Gemini. The company said OS-level features can capture chat content as it appears on screen, putting private data outside Signal's control. Signal is asking Apple, Google and Microsoft to disclose AI access mechanisms and give developers a clear opt-out.
An attacker exploited a vulnerability in Optimizely's web experimentation platform to plant malicious JavaScript on customer websites, hitting 178 organizations including BBC, CNN and Zoom. Help Net Security also reviewed Java4Shell, an OpenJDK Base64 decoder flaw, saying exploitation is tricky and depends on specific CPUs.
OpenRouter drops three providers after months of prompt theft
OpenRouter removed three AI inference providers and 125 endpoints from Zero Data Retention eligibility after Dragos found months of prompt theft. Researchers planted about two million prompts with unique web links and credentials from July to November and found stolen AWS keys, S3 bucket listings and Secrets Manager searches. OpenRouter said the breaches were at the providers, not in its own gateway.
Ghostscript patch fixes sandbox escape in common upload paths
A patched Ghostscript flaw, CVE-2026-125396, lets a crafted file escape the program's sandbox and run commands on a server that processes it. pwn.ai showed upload, preview and print features in about ten common applications could pass such files to Ghostscript through libraries such as ImageMagick, Pillow, libvips or Qt plugins. Artifex fixed the use-after-free in version 10.08.0 on September 8, with no reported attacks.